JP2020153743A - Ground fault factor estimation device, data generation device, ground fault factor estimation method, data generation method, and ground fault relay - Google Patents

Abstract

To provide a device and a method that are effective for improving reliability in estimating ground fault factors on the basis of ground fault current waveform data.SOLUTION: A ground fault factor estimation device 100 includes: a harmonic wave data calculation section 114 which calculates multivariable harmonic wave data indicating the intensity of a harmonic wave component in ground fault current waveform data, according to harmonic wave analysis of the ground fault current waveform data of a high-voltage rail track 91 where a ground fault has occurred; a position calculation section 115 which calculates a position in a data coordinate of the harmonic wave data, on the basis of coordinate definition data calculated in advance by applying decorrelation processing to multiple pieces of harmonic wave data, so that data coordinates are defined to perform mapping of the harmonic wave data; and a factor estimation section 116 which estimates a factor of the ground fault, on the basis of area data calculated in advance according to a correspondence relationship between the multiple pieces of harmonic wave data and the multiple pieces of factor information on the ground fault, and a position in the data coordinate of the harmonic wave data.SELECTED DRAWING: Figure 3

Description

The present disclosure relates to a ground fault factor estimation device, a data generation device, a ground fault factor estimation method, a data generation method, and a ground relay.

In Patent Document 1, at least the zero-phase voltage waveform and the zero-phase current waveform at the time of a ground fault of a distribution line are detected, data is recorded, waveform analysis is performed from multiple viewpoints, and the analysis values are input to a neural network. However, a method for estimating and outputting the cause of the accident is disclosed.

Japanese Unexamined Patent Publication No. 11-142466

An object of the present disclosure is to provide an apparatus and a method effective for improving the reliability of estimation of a ground fault factor based on ground fault current waveform data.

The ground fault factor estimation device according to one aspect of the present disclosure includes a current data acquisition unit that acquires ground fault current waveform data of a distribution line in which a ground fault has occurred, and a ground fault current by harmonic analysis of the ground fault current waveform data. Multiple harmonics so as to determine the harmonic data calculation unit that calculates the multivariate harmonic data indicating the intensity of the harmonic component in the waveform data and the data coordinates for mapping the harmonic data of the ground fault current waveform data. To show the relationship between the position calculation unit that calculates the position of the harmonic data in the data coordinates based on the coordinate definition data calculated in advance by the uncorrelation processing for the wave data, and the position in the data coordinates and the cause of the ground fault. In addition, a factor that estimates the cause of the ground fault based on the area data calculated in advance based on the correspondence between the multiple harmonic data and the factor information of the multiple ground faults and the position of the harmonic data in the data coordinates. It is equipped with an estimation unit.

The ground fault factor estimation device calculates coordinate definition data by a data storage unit that stores a plurality of harmonic data in association with each other and a plurality of ground fault factor information, and by uncorrelated processing for the plurality of harmonic data. A data coordinate calculation unit for calculating data coordinates and an area data calculation unit for calculating area data based on the correspondence between a plurality of harmonic data and a plurality of ground fault factor information may be further provided.

The data coordinate calculation unit may calculate coordinate definition data that determines data coordinates in two to four dimensions.

The data coordinate calculation unit may calculate the coordinate definition data by principal component analysis.

The area data calculation unit may calculate the area data by the support vector machine.

The ground fault factor estimation device adds an actual data acquisition unit that acquires actual factor information of the ground fault and update data that associates the harmonic data corresponding to the ground fault with the actual factor information to the data storage unit. Based on the actual data addition section and the harmonic data of the update data, the data coordinate calculation section updates the coordinate definition data, and based on the correspondence between the harmonic data in the update data and the cause information of the ground fault. The area data calculation unit may further include an update unit for updating the area data.

The data generator for estimating ground fault factors according to other aspects of the present disclosure includes multivariate harmonic data showing the intensity of harmonic components in the ground fault current waveform data of the distribution line in which the ground fault has occurred, and ground faults. The data storage unit that stores the factor information in association with the above and the coordinate definition data that determines the data coordinates for mapping the harmonic data are calculated by uncorrelation processing for the multiple harmonic data stored by the data storage unit. Area data that calculates the area data showing the relationship between the position in the data coordinates and the cause of the ground fault based on the correspondence between the data coordinate calculation unit and the multiple harmonic data and the multiple factor information in the data storage unit. It is equipped with a calculation unit.

The data generator has a simulated data acquisition unit that acquires ground fault current waveform data and ground fault factor information from a simulated facility that reproduces a ground fault with known factors, and a ground fault current waveform acquired by the simulated data acquisition unit. It may further include a harmonic data calculation unit that calculates harmonic data by harmonic analysis of the data.

The ground fault factor estimation method according to still another aspect of the present disclosure is a ground fault current waveform by acquiring ground fault current waveform data of a distribution line in which a ground fault has occurred and by harmonic analysis of the ground fault current waveform data. Calculated in advance by calculating multivariate harmonic data that indicates the intensity of the harmonic component in the data, and by uncorrelation processing for multiple harmonic data so as to determine the data coordinates for mapping the harmonic data. To calculate the position of the harmonic data in the data coordinates based on the coordinate definition data, and to show the relationship between the position in the data coordinates and the cause of the ground fault, the multiple harmonic data and the multiple ground faults It includes estimating the cause of the ground fault based on the area data calculated in advance based on the correspondence with the factor information and the position of the harmonic data in the data coordinates.

The ground fault factor estimation method is to accumulate multiple harmonic data and multiple ground fault factor information in association with each other, and to calculate coordinate definition data by uncorrelated processing for the multiple harmonic data. And the calculation of the area data based on the correspondence between the plurality of harmonic data and the factor information of the plurality of ground faults may be further included.

The data generation method for estimating the ground fault factor according to still another aspect of the present disclosure includes multivariate harmonic data showing the intensity of the harmonic component in the ground fault current waveform data of the distribution line in which the ground fault has occurred, and ground. Accumulation of related factor information in association with each other, and calculation of coordinate definition data for defining data coordinates for mapping harmonic data by uncorrelation processing for a plurality of accumulated harmonic data. This includes calculating area data showing the relationship between the position in the data coordinates and the factor of the ground fault based on the correspondence between the harmonic data of the above and a plurality of factor information.

The data generation method is to acquire the ground fault current waveform data and the ground fault factor information from the simulated equipment that reproduces the ground fault with known factors, and to analyze the harmonic current waveform data acquired from the simulated equipment. It may further include the calculation of harmonic data.

The ground fault relay according to still another aspect of the present disclosure is a ground fault current based on a current data acquisition unit that acquires ground fault current waveform data of a distribution line in which a ground fault has occurred and a harmonic analysis of the ground fault current waveform data. Multiple harmonics so as to determine the harmonic data calculation unit that calculates the multivariate harmonic data indicating the intensity of the harmonic component in the waveform data and the data coordinates for mapping the harmonic data of the ground fault current waveform data. Show the relationship between the position calculation unit that calculates the position of the harmonic data in the data coordinates and the position in the data coordinates and the cause of the ground fault based on the coordinate definition data calculated in advance by the uncorrelation processing for the wave data. In addition, a factor that estimates the cause of the ground fault based on the area data calculated in advance based on the correspondence between the multiple harmonic data and the factor information of the multiple ground faults and the position of the harmonic data in the data coordinates. It includes an estimation unit.

According to the present disclosure, it is possible to provide an apparatus and a method effective for improving the reliability of estimating the ground fault factor based on the ground fault current waveform data.

It is a schematic diagram which shows the schematic structure of the ground fault factor estimation system. It is a graph which illustrates the current waveform at the time of the ground fault occurrence. It is a block diagram which illustrates the functional configuration of the learning server. It is a block diagram which illustrates the functional structure of the ground fault factor estimation apparatus. It is a block diagram which illustrates the hardware configuration of the learning server and the ground fault factor estimation device. It is a flowchart which illustrates the data generation procedure for the ground fault factor estimation. It is a graph which illustrates the data generated by the procedure of FIG. It is a flowchart which illustrates the data generation procedure for the ground fault factor estimation. It is a flowchart which exemplifies the ground fault factor estimation procedure. It is a graph for showing the estimation example of the ground fault factor.

Hereinafter, embodiments will be described in detail with reference to the drawings. In the description, the same elements or elements having the same function are designated by the same reference numerals, and duplicate description will be omitted.

[Ground fault factor estimation system]
The ground fault factor estimation system shown in FIG. 1 is a system for estimating the ground fault factor based on the current waveform of the distribution line when the ground fault of the distribution line occurs. The ground fault here also includes a micro ground fault in which the insulation resistance is reduced. The distribution line may be any line that transmits power from a power source facility such as a power plant to a consumer facility, and there is no limitation on the type. Specific examples of the distribution line include a high voltage line of 6600V.

As an example, the ground fault factor estimation system 1 estimates the cause of a ground fault that occurs on a railroad track in a high-voltage consumer facility. The high-voltage line 91 includes a switch 11 for disconnecting the consumer equipment from the high-voltage line 91 when a ground fault or the like occurs, and a relay 12 for opening and closing the switch 11 based on the current information of the high-voltage line 91. Ground relay) is provided. Therefore, it is possible to acquire the ground fault current waveform data from the relay 12 when a ground fault occurs in the line in the high-voltage consumer equipment. Therefore, the ground fault factor estimation system 1 estimates the ground fault factor based on the ground fault current waveform data acquired from the relay 12.

For example, the ground fault factor estimation system 1 has a ground fault factor estimation device 100 and a learning server 200. The ground fault factor estimation device 100 and the learning server 200 are connected via a network line 92. The network line 92 may be a local area network or a wide area network (for example, the Internet). The ground fault factor estimation device 100 and the learning server 200 may be connected by a communication cable without going through the network line 92, or may be connected by short-range wireless communication. The learning server 200 generates estimation data for estimating the ground fault factor, and the ground fault factor estimation device 100 includes the estimation data generated by the learning server 200 and the ground fault current waveform data acquired from the relay 12. Estimate the ground fault factor based on. Hereinafter, the configurations of the learning server 200 and the ground fault factor estimation device 100 will be specifically illustrated.

(Learning server)
The learning server 200 accumulates the multivariable harmonic data indicating the intensity of the harmonic component in the ground fault current waveform data of the high-pressure line 91 in which the ground fault has occurred, and the cause information of the ground fault in association with each other. To calculate the coordinate definition data that defines the data coordinates for mapping the harmonic data by uncorrelation processing for the accumulated multiple harmonic data, and to the correspondence between the multiple harmonic data and the multiple factor information. Based on this, it is configured to calculate and execute area data indicating the relationship between the position in the data coordinates and the cause of the ground fault. The learning server 200 may be configured to acquire data for storage from the ground fault simulation facility 20 that reproduces the ground fault by a known factor. Specifically, the learning server 200 acquires ground fault current waveform data and ground fault factor information from the ground fault simulation facility 20, and harmonics of the ground fault current waveform data acquired from the ground fault simulation facility 20. It may be configured to further perform the calculation of harmonic data by analysis.

For example, the ground fault simulation facility 20 includes an artificial ground fault generator 21, a simulated switch 22, and a simulated relay 23. The artificial ground fault generator 21 is a device that artificially causes a ground fault in the high-voltage line 91 due to a plurality of types of factors assumed in actual power distribution equipment. The simulated switch 22 disconnects the ground fault simulating equipment 20 from the high-voltage line 91 in response to the occurrence of a ground fault in the artificial ground fault generator 21. The simulated relay 23 opens and closes the simulated switch 22 based on information on the current in the high-voltage line 91 and the like. When the ground fault simulation equipment 20 is configured in this way, the learning server 200 may acquire ground fault current waveform data when the artificial ground fault generator 21 is generating a ground fault from the simulated relay 23. ..

Specific examples of the factors of the ground fault that can be simulated by the ground fault simulation facility 20 include cable deterioration, gap discharge, insulator cracking, insulator fouling, metal contact, bird and beast contact, and the like. FIG. 2 is a graph illustrating the current waveform for each factor of the ground fault. In each graph, the horizontal axis shows the elapsed time and the vertical axis shows the current. FIG. 2A is a current waveform when a ground fault occurs due to cable deterioration. FIG. 2B is a current waveform when a ground fault occurs due to gap discharge. FIG. 2C is a current waveform when a ground fault occurs due to insulator cracking. FIG. 2D is a current waveform when a ground fault occurs due to insulator fouling. FIG. 2 (e) is a current waveform when a ground fault occurs due to metal contact. FIG. 2F is a current waveform when a ground fault occurs due to contact with birds and beasts. As illustrated in FIG. 2, the current waveform depends on the cause of the ground fault. However, even if the factors are the same, the current waveform changes due to various factors, so it is not easy to estimate the cause of the ground fault from the current waveform itself. Therefore, the learning server 200 performs conversion processing on the ground fault current waveform data so that the characteristics of each ground fault factor can be easily displayed, and performs machine learning based on the data after the conversion processing to obtain data for estimating the ground fault factor. Derived.

As illustrated in FIG. 3, the learning server 200 has a data storage unit 211, a simulated data acquisition unit 212, a harmonic data calculation unit 213, and a functional configuration (hereinafter referred to as a “functional module”). It has a simulated data addition unit 214, a data coordinate calculation unit 215, an area data calculation unit 216, an estimation data storage unit 217, and an estimation data transmission unit 218.

The data storage unit 211 stores a plurality of harmonic data and a plurality of ground fault factor information in association with each other. The simulated data acquisition unit 212 acquires ground fault current waveform data and ground fault factor information (label data) from the ground fault simulation equipment 20. As described above, the simulated data acquisition unit 212 may acquire ground fault current waveform data from the simulated relay 23. Further, the simulated data acquisition unit 212 may acquire the cause information of the ground fault from the artificial ground fault generator 21. Acquiring the cause information of the ground fault from the ground fault simulation facility 20 also includes acquiring the cause information of the ground fault from the worker of the ground fault simulation facility 20. For example, the simulated data acquisition unit 212 may acquire the cause information of the ground fault from the terminal device used by the worker of the ground fault simulation equipment 20 via the network line, or the worker of the ground fault simulation equipment 20 learns. The cause information of the ground fault may be acquired based on the direct input to the server 200 (for example, the input to the input device 297).

The harmonic data calculation unit 213 calculates the harmonic data by harmonic analysis of the ground fault current waveform data acquired by the simulated data acquisition unit 212. As described above, the harmonic data is multivariable data indicating the intensity of the harmonic component in the ground fault current waveform data. "Multivariable" means to include two or more variables. Specific examples of the variables included in the harmonic data include the intensities of harmonic components (for example, current amplitude) at two or more orders. For example, the harmonic data calculation unit 213 may calculate the current amplitude for each frequency by a fast Fourier transform or the like, and extract the harmonic component from the calculation result. The harmonic component is a component having a frequency that is an integral multiple of the frequency of the fundamental wave component. The method for calculating harmonic data is not limited to the Fourier transform. For example, the harmonic data calculation unit 213 may calculate the harmonic data by digital filtering corresponding to a bandpass filter for each harmonic order (multiple of the frequency of the fundamental wave component).

The simulated data addition unit 214 adds the harmonic data calculated by the harmonic data calculation unit 213 and the ground fault factor information acquired by the simulated data acquisition unit 212 to the data storage unit 211 in association with each other.

The data coordinate calculation unit 215 calculates the coordinate definition data by uncorrelated processing on the plurality of harmonic data accumulated by the data storage unit 211. As described above, the coordinate definition data is data that determines the data coordinates for mapping the harmonic data. For example, the coordinate definition data includes information corresponding to the origin and information indicating the directions of a plurality of coordinate axes. As a specific example of mapping, high frequency data may be represented by a position vector in a data coordinate system.

The uncorrelated processing means a process for reducing the dimension of data by reducing a plurality of variables to a smaller number of variables having a low correlation with each other by integrating variables having a high correlation with each other. Specific examples of the known uncorrelated processing include principal component analysis and the like. By the uncorrelated processing, the data coordinate calculation unit 215 may calculate the coordinate definition data that determines the data coordinates in 2 to 4 dimensions.

For example, the data coordinate calculation unit 215 calculates the coordinate definition data by principal component analysis. For example, the data coordinate calculation unit 215 calculates the eigenvalues of the covariance matrix of the plurality of harmonic data. Each eigenvalue has an eigenvector, and the eigenvectors for different eigenvalues are orthogonal to each other. The data coordinate calculation unit 215 selects as coordinate definition data in order from the eigenvector having the largest eigenvalue until the cumulative contribution rate exceeds a predetermined value (for example, 90%). The coordinate axes of the data coordinates are determined by the plurality of eigenvectors selected in this way.

The area data calculation unit 216 calculates the area data based on the correspondence between the plurality of harmonic data in the data storage unit 211 and the factor information of the plurality of ground faults. As described above, the area data is data showing the relationship between the position in the data coordinates and the factor of the ground fault. For example, the area data calculation unit 216 converts a plurality of harmonic data into a plurality of position vectors in the data coordinates. Hereinafter, a point whose position is determined by each position vector is referred to as a “data point”.

The area data calculation unit 216 associates the ground fault factor information with each data point based on the correspondence relationship between the plurality of harmonic data and the plurality of ground fault factor information. As a result, the distribution information of the data points for each factor of the ground fault is calculated. The distribution information of the data points for each ground fault factor shows the relationship between the position in the data coordinates and the ground fault factor. Therefore, the distribution information of the data points for each ground fault factor corresponds to an example of the above area data.

The area data calculation unit 216 may calculate the boundary data that defines the distribution area of the data points for each ground fault factor as the area data. For example, the area data calculation unit 216 may calculate boundary data by a support vector machine.

The estimation data storage unit 217 stores the coordinate definition data calculated by the data coordinate calculation unit 215 and the area data calculated by the area data calculation unit 216 as data for estimating the ground fault factor. The data for estimating the ground fault factor stored in the estimation data storage unit 217 may further include a plurality of harmonic data used for calculating the coordinate definition data and the area data.

The estimation data transmission unit 218 transmits the ground fault factor estimation data stored in the estimation data storage unit 217 to the ground fault factor estimation device 100. The learning server 200 acquires the actual factor information of the new ground fault, and acquires the update data in which the harmonic data corresponding to the new ground fault and the actual factor information of the new ground fault are associated with each other. Further accumulation, updating the coordinate definition data based on the harmonic data of the update data, and updating the area data based on the correspondence between the harmonic data in the update data and the cause information of the ground fault. , May be configured to perform further. For example, the learning server 200 has a performance data acquisition unit 221, a performance data addition unit 222, and an update unit 223 as functional modules.

The actual data acquisition unit 221 acquires the harmonic data corresponding to the new ground fault and the actual factor information of the new ground fault from the ground fault factor estimation device 100. The actual factor information is input to the ground fault factor estimation device 100 by the operator of the repair work for eliminating the ground fault. The actual data acquisition unit 221 may acquire the actual factor information from the terminal of the worker of the repair work or the like without going through the ground fault factor estimation device 100.

The actual data addition unit 222 adds update data in which the actual data acquired by the actual data acquisition unit 221 and the factor information are associated with each other to the data storage unit 211. The update unit 223 causes the data coordinate calculation unit 215 to update the coordinate definition data based on the harmonic data of the update data, and the area is based on the correspondence between the harmonic data in the update data and the cause information of the ground fault. Have the data calculation unit 216 update the area data.

(Ground fault factor estimation device)
The ground fault factor estimation device 100 acquires the ground fault current waveform data of the high-pressure line 91 in which a new ground fault of unknown factor has occurred, and the harmonic analysis of the ground fault current waveform data causes the ground fault current waveform data. It was calculated in advance by calculating the multivariable harmonic data indicating the intensity of the harmonic component in the above and by uncorrelation processing for a plurality of harmonic data so as to determine the data coordinates for mapping the harmonic data. Multiple harmonic data and multiple ground fault factors to show the relationship between calculating the position of the harmonic data in the data coordinates based on the coordinate definition data and the relationship between the position in the data coordinates and the ground fault factor. It is configured to perform estimation of the cause of a new ground fault based on the area data calculated in advance based on the correspondence with the information and the position of the harmonic data in the data coordinates. As illustrated in FIG. 4, for example, the ground fault factor estimation device 100 has, as functional modules, an estimation data receiving unit 111, an estimation data storage unit 112, a current data acquisition unit 113, and a harmonic data calculation unit 114. It has a position calculation unit 115, a factor estimation unit 116, and an estimation result display unit 117.

The estimation data receiving unit 111 receives the above-mentioned ground fault factor estimation data from the estimation data transmitting unit 218. The estimation data storage unit 112 stores the ground fault factor estimation data received by the estimation data receiving unit 111. The current data acquisition unit 113 acquires the ground fault current waveform data of the high-voltage line 91 in which a new ground fault of unknown cause has occurred from the relay 12. The harmonic data calculation unit 114 calculates the harmonic data by harmonic analysis of the ground fault current waveform data acquired by the current data acquisition unit 113. The specific content of the harmonic analysis performed by the harmonic data calculation unit 114 is the same as the specific content of the harmonic analysis performed by the harmonic data calculation unit 213.

The position calculation unit 115 calculates the position of the harmonic data calculated by the harmonic data calculation unit 114 in the above data coordinates based on the coordinate definition data stored in the estimation data storage unit 112. The position calculation unit 115 corrects the coordinate definition data based on the harmonic data calculated by the harmonic data calculation unit 114, and calculates the position of the harmonic data in the data coordinates determined by the corrected coordinate definition data. Good. For example, the position calculation unit 115 uses the harmonic data calculated by the harmonic data calculation unit 114 when the ground fault factor estimation data stored in the estimation data storage unit 112 includes a plurality of harmonic data. In addition to the plurality of harmonic data, the coordinate definition data is recalculated, and the position of the harmonic data in the data coordinates determined by the recalculated coordinate definition data is calculated.

The factor estimation unit 116 estimates a new ground fault factor based on the area data stored in the estimation data storage unit 112 and the position of the harmonic data calculated by the position calculation unit 115 in the data coordinates. The estimation result display unit 117 displays the estimation result by the factor estimation unit 116. For example, the estimation result display unit 117 displays the estimation result by the factor estimation unit 116 on a display device 196 (described later) or the like.

(Hardware configuration of learning server and ground fault factor estimation device)
The learning server 200 is a computer having a communication function, and includes a circuit 290 as shown in FIG. The circuit 290 includes a processor 291, a memory 292, a storage 293, an input / output port 294, a communication port 295, a display device 296, and an input device 297. The storage 293 is a non-volatile storage medium (eg, flash memory) that can be read by a computer. The storage 293 accumulates the harmonic data in association with the factor information of the ground fault, calculates coordinate definition data by uncorrelated processing on the accumulated plurality of harmonic data, and performs a plurality of harmonics. It stores a program that causes the learning server 200 to calculate the area data based on the correspondence between the wave data and the plurality of factor information. For example, the storage 293 stores a program for configuring the functional module of the learning server 200 described above.

The memory 292 temporarily stores the program loaded from the storage 293, the calculation result by the processor 291 and the like. The processor 291 constitutes each functional module of the learning server 200 by executing the above program in cooperation with the memory 292. The input / output port 294 inputs / outputs an electric signal to / from the simulated relay 23 in response to a command from the processor 291. The communication port 295 performs network communication with the ground fault factor estimation device 100 via the network line 92 in response to a command from the processor 291.

The display device 296 and the input device 297 function as a user interface of the learning server 200. The display device 296 includes, for example, a liquid crystal monitor and is used for displaying information to the user. The input device 297 is, for example, a keyboard or the like, and acquires input information by the user. The display device 296 and the input device 297 may be integrated as a so-called touch panel.

The ground fault factor estimation device 100 is a computer having a communication function, and includes a circuit 190. The circuit 190 includes a processor 191 and a memory 192, a storage 193, an input / output port 194, a communication port 195, a display device 196, and an input device 197. The storage 193 is a non-volatile storage medium (eg, flash memory) that can be read by a computer. The storage 193 is based on the acquisition of the ground fault current waveform data of the high-pressure line 91 in which a new ground fault of unknown cause has occurred, the calculation of the harmonic data of the ground fault current waveform data, and the coordinate definition data. The ground fault factor estimation device 100 calculates the position of the harmonic data in the data coordinates and estimates a new ground fault factor based on the area data and the position of the harmonic data in the data coordinates. Memorizes the program to be executed by. For example, the storage 193 stores a program for configuring the functional module of the ground fault factor estimation device 100 described above.

The memory 192 temporarily stores the program loaded from the storage 193, the calculation result by the processor 191 and the like. The processor 191 constitutes each functional module of the ground fault factor estimation device 100 by executing the above program in cooperation with the memory 192. The input / output port 194 inputs / outputs an electric signal to / from the relay 12 in response to a command from the processor 191. The communication port 195 performs network communication with the learning server 200 via the network line 92 in response to a command from the processor 191.

The display device 196 and the input device 197 function as a user interface of the ground fault factor estimation device 100. The display device 196 includes, for example, a liquid crystal monitor and is used for displaying information to the user. The input device 197 is, for example, a keyboard or the like, and acquires input information by the user. The display device 196 and the input device 197 may be integrated as a so-called touch panel.

The configuration of the ground fault factor estimation system 1 illustrated above is just an example and can be changed as appropriate. For example, the learning server 200 does not necessarily have to acquire the data for storage from the ground fault simulation facility 20, and is configured to calculate the coordinate definition data and the area data based only on the actual data in the consumer facility. May be good.

The ground fault factor estimation device 100 may be incorporated in the relay 12. Further, both the ground fault factor estimation device 100 and the learning server 200 may be incorporated in the relay 12.

The ground fault factor estimation device 100 may be configured to transmit the ground fault factor estimation result to a terminal device such as a manager of a consumer facility via a network line. In this case, the display by the estimation result display unit 117 described above can be omitted. Specific examples of the terminal device include mobile terminals such as laptop computers, tablet computers, and smartphones. The terminal device may be a stationary terminal such as a desktop computer.

The ground fault factor estimation device 100 may be incorporated in a terminal device such as a manager of consumer equipment. In this case, the ground fault factor estimation system 1 may further include a data relay device for transmitting ground fault current waveform data from the relay 12 to the terminal device by short-range wireless communication or the like. This relay device may be incorporated in the relay 12.

The ground fault factor estimation device 100 may be configured to transmit the ground fault factor estimation result to the mobile terminal by a communication cable, short-range wireless communication, or the like. Further, the ground fault factor estimation device 100 may be connected to the learning server 200 via the mobile terminal.

At least a part of the functions of the ground fault factor estimation device 100 may be incorporated in the learning server 200. For example, the learning server 200 calculates the position of the harmonic data in the data coordinates based on the coordinate definition data, and estimates a new ground fault factor based on the area data and the position of the harmonic data in the data coordinates. And may be configured to do more. In this case, the ground fault factor estimation device 100 transmits the calculation result of the harmonic data to the learning server 200, and the ground fault factor estimation result based on the position of the harmonic data in the data coordinates and the area data is transmitted to the learning server 200. Will be received from.

Further, the learning server 200 further executes the acquisition of the ground fault current waveform data of the high-pressure line 91 in which a new ground fault of unknown cause has occurred, and the calculation of the harmonic data of the ground fault current waveform data. It may be configured as follows. In this case, the ground fault factor estimation device 100, which is separate from the learning server 200, is not required, but the device that transmits ground fault current waveform data from the relay 12 to the learning server 200 and the ground fault factor estimation result in the learning server 200 can be obtained. A device to notify is needed. Therefore, the ground fault factor estimation system 1 acquires ground fault current waveform data from the relay 12 and transmits the ground fault current waveform data to the learning server 200 via the network line 92 (hereinafter referred to as “waveform data relay device”) and a learning server. Even if it is further equipped with a relay device (hereinafter, referred to as "estimation result relay device") that acquires the ground fault factor estimation result from 200 and transmits it to a terminal device such as a manager of consumer equipment via a network line 92. Good. The waveform data relay device may be incorporated in the relay 12 or may be incorporated in the terminal device. The estimation result relay device may be incorporated in the learning server 200.

[Ground fault factor estimation method]
Subsequently, as an example of the ground fault factor estimation method, the ground fault factor estimation procedure executed by the ground fault factor estimation system 1 will be illustrated. This procedure includes a data generation procedure for estimating the ground fault factor by the learning server 200 and a ground fault factor estimation procedure by the ground fault factor estimation device 100. Hereinafter, each procedure will be illustrated in detail.

(Data generation procedure for estimating ground fault factors)
The data generation procedure for estimating the ground fault factor calculates the coordinate definition data by accumulating the harmonic data and the ground fault factor information in association with each other and by uncorrelated processing for the accumulated multiple harmonic data. This includes calculating area data based on the correspondence between a plurality of harmonic data and a plurality of factor information. In this procedure, harmonic data is obtained by acquiring ground fault current waveform data and ground fault factor information from the ground fault simulation facility 20 and by harmonic analysis of the ground fault current waveform data acquired from the ground fault simulation facility 20. , And the calculated harmonic data may be stored in association with the factor information acquired from the ground fault simulation facility 20.

As illustrated in FIG. 6, the learning server 200 first executes steps S01, S02, and S03. In step S01, the simulated data acquisition unit 212 waits for the ground fault simulation facility 20 to start reproducing the ground fault due to a known factor. In step S02, the simulated data acquisition unit 212 acquires the ground fault current waveform data and the ground fault factor information from the ground fault simulation facility 20. In step S03, the harmonic data calculation unit 213 calculates the harmonic data by harmonic analysis of the ground fault current waveform data acquired by the simulated data acquisition unit 212. The harmonic data calculation unit 213 may calculate the harmonic data after performing low-pass type filtering on the ground fault current waveform data and removing unnecessary high frequency components.

FIG. 7A exemplifies the ground fault current waveform data before filtering, and FIG. 7B exemplifies the ground fault current waveform data after filtering. In each graph, the horizontal axis shows the elapsed time and the vertical axis shows the current. FIG. 7C is a graph illustrating harmonic data, and includes the intensities (current amplitudes) of harmonic components from the first order to the fifteenth order from the left. That is, the harmonic data illustrated in FIG. 7 (c) is the harmonic data of 15 variables.

Returning to FIG. 6, the learning server 200 then executes steps S04 and S05. In step S04, the simulated data addition unit 214 adds the harmonic data calculated by the harmonic data calculation unit 213 to the ground fault factor information acquired by the simulated data acquisition unit 212 in association with the data storage unit 211. .. The simulated data addition unit 214 may exclude the correspondence between the harmonic data and the ground fault factor information from the storage target in the data storage unit 211 when the correspondence relationship is clearly different from the past tendency. .. In step S05, the data coordinate calculation unit 215 confirms whether or not the number of data stored in the data storage unit 211 has reached a preset learning threshold value. The learning threshold is preset so that the accuracy of ground fault factor estimation is at a desired level.

If it is determined in step S05 that the number of data stored in the data storage unit 211 has not reached the learning threshold value, the learning server 200 returns the process to step S01. After that, until the number of data stored in the data storage unit 211 reaches the learning threshold, the accumulation of harmonic data and ground fault factor information is repeated.

When it is determined in step S05 that the number of data stored in the data storage unit 211 has reached the learning threshold value, the learning server executes steps S06 and S07. In step S06, the data coordinate calculation unit 215 calculates the coordinate definition data by uncorrelated processing on the plurality of harmonic data accumulated by the data storage unit 211, and saves the calculation result in the estimation data storage unit 217. .. In step S07, the area data calculation unit 216 calculates the area data based on the correspondence between the plurality of harmonic data in the data storage unit 211 and the factor information of the plurality of ground faults, and stores the calculation result as data for estimation. Store in part 217.

FIG. 7D is a graph illustrating the data coordinates determined by the coordinate definition data calculated by the data coordinate calculation unit 215 and the area data calculated by the area data calculation unit 216. In this example, two-dimensional data coordinates are determined by principal component analysis. The horizontal axis of the graph is the first coordinate axis (first principal component of the principal component analysis), and the vertical axis of the graph is the second coordinate axis (second principal component of the principal component analysis). The square marks plotted in the graph indicate the data points where the cause of the ground fault is cable deterioration. The triangles plotted in the graph indicate data points where the cause of the ground fault is other than cable deterioration. The boundary line PL in the graph indicates the boundary between the distribution area of data points whose ground fault factor is cable deterioration and the distribution area of data points whose ground fault factor is other than cable deterioration. In this example, the distribution area of the data points whose cause of the ground fault is cable deterioration is surrounded by the boundary line PL. The boundary data that determines the boundary line PL is calculated by the support vector machine.

Note that FIG. 7D shows an example in which the area data divides the data coordinates into an area of cable deterioration and an area other than cable deterioration, but the present invention is not limited to this. For example, it is possible to calculate the area data so as to further subdivide the area other than the cable deterioration.

As shown in FIG. 8, the learning server 200 then executes step S11. In step S11, the estimation data transmission unit 218 confirms whether or not the transmission of the ground fault factor estimation data is requested by the ground fault factor estimation device 100. When it is determined in step S11 that the transmission of data for estimating the ground fault factor is requested, the learning server 200 executes step S12. In step S12, the estimation data transmission unit 218 transmits the ground fault factor estimation data stored in the estimation data storage unit 217 to the ground fault factor estimation device 100. After that, the ground fault factor estimation device 100 returns the process to step S11. After that, the data for estimating the ground fault factor is repeatedly transmitted in response to the request from the ground fault factor estimation device 100.

When it is determined in step S11 that the transmission of the data for estimating the ground fault factor is not requested, the learning server 200 executes step S13. In step S13, the actual data acquisition unit 221 confirms whether or not the harmonic data corresponding to the new ground fault and the actual factor information of the new ground fault are transmitted from the ground fault factor estimation device 100. If it is determined in step S13 that the harmonic data and the actual factor information have not been transmitted, the learning server 200 returns the process to step S11.

When it is determined in step S13 that the harmonic data and the actual factor information are transmitted, the learning server 200 executes steps S14 and S15. In step S14, the actual data addition unit 222 adds the update data in which the actual data acquired by the actual data acquisition unit 221 and the factor information are associated with each other to the data storage unit 211. The actual data addition unit 222 may exclude these from the addition target to the data storage unit 211 when the correspondence between the actual data and the factor information is clearly different from the past tendency. In step S15, the update unit 223 confirms whether or not the number of update data added to the data storage unit 211 has reached a predetermined update threshold value. The update threshold is set in advance so as to improve the accuracy of ground fault factor estimation by updating to a desired level. When it is determined in step S15 that the number of update data has not reached the update threshold value, the learning server 200 returns the process to step S11.

When it is determined in step S15 that the number of update data has reached the update threshold value, the ground fault simulation facility 20 executes steps S16 and S17. In step S16, the update unit 223 causes the data coordinate calculation unit 215 to update the coordinate definition data based on the harmonic data of the update data. The data coordinate calculation unit 215 recalculates the coordinate definition data based on both the plurality of harmonic data accumulated up to step S06 and the plurality of harmonic data added up to step S16, and the calculation result. Is overwritten in the estimation data storage unit 217.

In step S17, the update unit 223 causes the area data calculation unit 216 to update the area data based on the correspondence between the harmonic data in the update data and the cause information of the ground fault. The area data calculation unit 216 recalculates the area data based on the above-mentioned correspondence between the data accumulated up to step S06 and the data added up to step S16, and stores the calculation result as data for estimation. Overwrite part 217. After that, the learning server 200 returns the process to step S01. After that, the update of the data for ground fault estimation based on the actual data is repeated.

(Ground fault factor estimation procedure)
The ground fault factor estimation procedure is to acquire the ground fault current waveform data of the distribution line where a new ground fault of unknown factor has occurred, and to calculate the above harmonic data by harmonic analysis of the ground fault current waveform data. To calculate the position of the harmonic data in the data coordinates based on the coordinate definition data, and to estimate the cause of the new ground fault based on the area data and the position of the harmonic data in the data coordinates. including.

As illustrated in FIG. 9, the ground fault factor estimation device 100 first executes steps S21, S22, and S23. In step S21, the estimation data receiving unit 111 requests the learning server 200 to transmit data for estimating the ground fault factor. In step S22, the estimation data receiving unit 111 receives the ground fault factor estimation data transmitted from the estimation data transmission unit 218 of the learning server 200 and stores it in the estimation data storage unit 112. In step S23, the current data acquisition unit 113 confirms whether or not a ground fault has been detected in the relay 12.

If it is determined in step S23 that a ground fault has been detected, the ground fault factor estimation device 100 executes steps S24, S25, S26, and S27. In step S24, the current data acquisition unit 113 acquires the ground fault current waveform data of the high-voltage line 91 in which the ground fault has occurred from the relay 12. In step S25, the harmonic data calculation unit 114 calculates the harmonic data by harmonic analysis of the ground fault current waveform data acquired by the current data acquisition unit 113. In step S26, the position calculation unit 115 calculates the position of the harmonic data calculated by the harmonic data calculation unit 114 in the above data coordinates based on the coordinate definition data stored in the estimation data storage unit 112. In step S27, the factor estimation unit 116 estimates a new ground fault factor based on the area data stored in the estimation data storage unit 112 and the position of the harmonic data calculated by the position calculation unit 115 in the data coordinates. To do.

FIG. 10 is a graph for showing an estimation example of ground fault factors. The boundary line PL and points indicating the positions calculated by the position calculation unit 115 (hereinafter, referred to as “estimated data points”) are shown in the data coordinates illustrated in FIG. 7 (d). Since the square data points are located inside the boundary line PL, it is presumed that the cause of the ground fault corresponding to the data points is the deterioration of the cable. Since the triangular data points are located outside the boundary line PL, it is presumed that the cause of the ground fault corresponding to the data points is other than the deterioration of the cable.

Returning to FIG. 9, the ground fault factor estimation device 100 then executes step S28. In step S28, the estimation result display unit 117 displays the estimation result by the factor estimation unit 116. After that, the ground fault factor estimation device 100 returns the process to step S23. After that, the cause is estimated and displayed repeatedly according to the occurrence of the ground fault.

If it is determined in step S23 that no ground fault has been detected, the ground fault factor estimation device 100 executes step S31. In step S31, the estimation data receiving unit 111 confirms whether or not the data for estimating the ground fault factor of the estimation data storage unit 217 has been updated. If it is determined in step S31 that the data for estimating the ground fault factor has not been updated, the ground fault factor estimation device 100 returns the process to step S23.

If it is determined in step S31 that the data for estimating the ground fault factor has been updated, the ground fault factor estimation device 100 executes steps S32 and S33. In step S32, the estimation data receiving unit 111 requests the learning server 200 to transmit data for estimating the ground fault factor. In step S33, the estimation data receiving unit 111 receives the ground fault factor estimation data transmitted from the estimation data transmission unit 218 of the learning server 200, and overwrites the estimation data storage unit 112. After that, the ground fault factor estimation device 100 returns the process to step S23. After that, when the ground fault does not occur, the update of the data for estimating the ground fault factor is repeated.

[Effect of this embodiment]
As described above, the ground fault factor estimation device 100 includes a current data acquisition unit 113 that acquires ground fault current waveform data of a high-pressure line 91 in which a new ground fault of unknown cause has occurred, and ground fault current waveform data. To map the harmonic data of the ground fault current waveform data with the harmonic data calculation unit 114 that calculates the multivariate harmonic data indicating the intensity of the harmonic component in the ground fault current waveform data by the harmonic analysis of The position calculation unit 115 that calculates the position of the harmonic data in the data coordinates based on the coordinate definition data calculated in advance by the uncorrelation processing for the plurality of harmonic data so as to determine the data coordinates, and the position in the data coordinates. The area data calculated in advance based on the correspondence between the multiple harmonic data and the multiple ground fault factor information and the position of the harmonic data in the data coordinates so as to show the relationship between the and the ground fault factor. It is provided with a factor estimation unit 116 that estimates a new ground fault factor based on the above.

When a ground fault occurs, prompt restoration is required, but the content of restoration work and the number of personnel required for restoration work differ depending on the factors. Therefore, it is desirable to estimate the cause of the ground fault in advance with high reliability. On the other hand, according to this device, the harmonic data is calculated by the harmonic analysis of the ground fault current waveform data, and the data is based on the coordinate definition data calculated in advance by the uncorrelated processing for the plurality of harmonic data. The position of the harmonic data in the coordinates is calculated. Further, in order to show the relationship between the position in the data coordinates and the factor of the ground fault, the area data calculated in advance based on the correspondence relationship between the plurality of harmonic data and the factor information of the plurality of ground faults and the harmonic data The cause of the ground fault is estimated based on the position in the data coordinates of.

High-frequency data with similar characteristics tend to have the same ground fault factor. Further, in the above data coordinates determined by the uncorrelated processing for a plurality of harmonic data, the data points corresponding to the harmonic data are widely distributed, so that the difference in the distribution area according to the characteristics is likely to appear clearly. Therefore, in the above data coordinates, the relationship between the harmonic data and the cause of the ground fault can be easily clarified. Therefore, the ground fault estimation factor estimation device that estimates the ground fault factor based on the area data showing the relationship between the position in the data coordinates and the ground fault factor and the position in the data coordinates of the harmonic data is the ground fault factor. It is effective in improving the reliability of the estimation of.

Further, in the above data coordinates, it becomes easy to estimate the cause (for example, deterioration of equipment with time) even at the time of occurrence of a microground fault having a small feature appearing in harmonic data. If deterioration equipment and the like can be estimated at the time of occurrence of a microground fault, it will be possible to efficiently perform repair work before a short circuit occurs. It is also expected that this will prevent large-scale damage such as unexpected power outages caused by ground faults.

Further, in the above data coordinates, the relationship between the harmonic data and the cause of the ground fault can be easily clarified even by the number of data smaller than the number of data required for the backpropagation method such as a neural network. Therefore, it is particularly effective for estimating the cause of a ground fault on a high-voltage line, which has a low frequency of occurrence and is difficult to collect data.

The ground fault factor estimation device 100 includes a data storage unit 211 that stores a plurality of harmonic data and a plurality of ground fault factor information in association with each other, and coordinate definition data by uncorrelated processing for the plurality of harmonic data. The data coordinate calculation unit 215 for calculating the data, and the area data calculation unit 216 for calculating the area data based on the correspondence between the plurality of harmonic data and the factor information of the plurality of ground faults may be further provided.

The data coordinate calculation unit 215 may calculate coordinate definition data that determines data coordinates in two to four dimensions. In this case, the relationship between the harmonic data and the cause of the ground fault is further clarified. Therefore, it is more effective in improving the reliability of estimating the ground fault factor.

The data coordinate calculation unit 215 may calculate the coordinate definition data by principal component analysis. In this case, data coordinates that are more effective for clarifying the relationship between the harmonic data and the cause of the ground fault are determined. Therefore, it is more effective in improving the reliability of estimating the ground fault factor.

The area data calculation unit 216 may calculate the area data by the support vector machine. In this case, area data that more accurately shows the relationship between the position in the data coordinates and the factor of the ground fault is calculated. Therefore, it is more effective in improving the reliability of estimating the ground fault factor.

The ground fault factor estimation device 100 has an actual data acquisition unit 221 that acquires actual factor information of a new ground fault, and update data in which harmonic data corresponding to the new ground fault and actual factor information are associated with each other. The data coordinate calculation unit 215 updates the coordinate definition data based on the actual data addition unit 222 and the harmonic data of the update data, and the harmonic data in the update data and the cause of the ground fault. An update unit 223 that causes the area data calculation unit 216 to update the area data based on the correspondence with the information may be further provided. In this case, the coordinate definition data and the area data are brushed up by the feedback of the actual factor information. Therefore, it is more effective in improving the reliability of estimating the ground fault factor.

Although the embodiments have been described above, the present disclosure is not necessarily limited to the above-described embodiments, and various changes can be made without departing from the gist thereof.

100 ... Ground fault factor estimation device, 113 ... Current data acquisition unit, 114 ... Harmonic data calculation unit, 115 ... Position calculation unit, 116 ... Factor estimation unit, 211 ... Data storage unit, 212 ... Simulated data acquisition unit, 215 ... Data coordinate calculation unit, 216 ... Area data calculation unit, 221 ... Actual data acquisition unit, 222 ... Actual data addition unit, 223 ... Update unit.

Claims (13)

A current data acquisition unit that acquires ground fault current waveform data of a distribution line in which a ground fault has occurred,
A harmonic data calculation unit that calculates multivariate harmonic data indicating the intensity of harmonic components in the ground fault current waveform data by harmonic analysis of the ground fault current waveform data.
The harmonic data of the harmonic data is based on coordinate definition data calculated in advance by uncorrelated processing on a plurality of harmonic data so as to determine data coordinates for mapping the harmonic data of the ground fault current waveform data. A position calculation unit that calculates the position in the data coordinates,
Area data calculated in advance based on the correspondence between the plurality of harmonic data and the factor information of the plurality of ground faults and the harmonics so as to show the relationship between the position in the data coordinates and the factor of the ground fault. A ground fault factor estimation device including a factor estimation unit that estimates the cause of the ground fault based on the position of the wave data in the data coordinates. A data storage unit that stores the plurality of harmonic data in association with each other and the factor information of the plurality of ground faults.
A data coordinate calculation unit that calculates the coordinate definition data by the uncorrelated processing on the plurality of harmonic data,
The ground fault factor estimation device according to claim 1, further comprising an area data calculation unit that calculates the area data based on the correspondence between the plurality of harmonic data and the plurality of ground fault factor information. The ground fault factor estimation device according to claim 2, wherein the data coordinate calculation unit calculates the coordinate definition data that determines the data coordinates in two to four dimensions. The ground fault factor estimation device according to claim 2 or 3, wherein the data coordinate calculation unit calculates the coordinate definition data by principal component analysis. The ground fault factor estimation device according to any one of claims 2 to 4, wherein the area data calculation unit calculates area data by a support vector machine. The actual data acquisition unit that acquires the actual factor information of the ground fault,
An actual data addition unit that adds update data that associates the harmonic data corresponding to the ground fault with the actual factor information to the data storage unit, and
Based on the harmonic data of the update data, the data coordinate calculation unit updates the coordinate definition data, and based on the correspondence relationship between the harmonic data in the update data and the cause information of the ground fault. The ground fault factor estimation device according to any one of claims 2 to 5, further comprising an update unit that causes the area data calculation unit to update the area data. A data storage unit that stores multivariable harmonic data indicating the intensity of harmonic components in the ground fault current waveform data of a distribution line in which a ground fault has occurred, and the factor information of the ground fault in association with each other.
A data coordinate calculation unit that calculates coordinate definition data that determines data coordinates for mapping the harmonic data by uncorrelated processing on a plurality of harmonic data accumulated by the data storage unit.
An area data calculation unit that calculates area data indicating the relationship between the position in the data coordinates and the cause of the ground fault based on the correspondence between the plurality of harmonic data and the plurality of factor information in the data storage unit. A data generator for estimating ground fault factors, including. A simulated data acquisition unit that acquires the ground fault current waveform data and the ground fault factor information from a simulated facility that reproduces the ground fault with a known factor.
The data generation device according to claim 7, further comprising a harmonic data calculation unit that calculates the harmonic data by harmonic analysis of the ground fault current waveform data acquired by the simulated data acquisition unit. Acquiring the ground fault current waveform data of the distribution line where the ground fault occurred, and
By the harmonic analysis of the ground fault current waveform data, it is possible to calculate multivariate harmonic data indicating the intensity of the harmonic component in the ground fault current waveform data.
The position of the harmonic data in the data coordinates is calculated based on the coordinate definition data calculated in advance by the uncorrelated processing for the plurality of harmonic data so as to determine the data coordinates for mapping the harmonic data. To do and
Area data calculated in advance based on the correspondence between the plurality of harmonic data and the factor information of the plurality of ground faults and the harmonics so as to show the relationship between the position in the data coordinates and the factor of the ground fault. A method for estimating a ground fault factor, including estimating the factor of the ground fault based on the position of the wave data in the data coordinates. Accumulating the plurality of harmonic data and the factor information of the plurality of ground faults in association with each other.
To calculate the coordinate definition data by the uncorrelated processing on the plurality of harmonic data, and
The ground fault factor estimation method according to claim 9, further comprising calculating the area data based on the correspondence between the plurality of harmonic data and the plurality of ground fault factor information. Accumulation of multivariable harmonic data indicating the intensity of harmonic components in the ground fault current waveform data of a distribution line in which a ground fault has occurred and the factor information of the ground fault in association with each other.
Coordinate definition data that determines the data coordinates for mapping the harmonic data is calculated by uncorrelated processing on the accumulated plurality of harmonic data, and
For estimating ground fault factors, including calculating area data indicating the relationship between the position at the data coordinates and the ground fault factor based on the correspondence between the plurality of harmonic data and the plurality of factor information. Data generation method. Acquiring the ground fault current waveform data and the ground fault factor information from a simulated facility that reproduces the ground fault with known factors, and
The data generation method according to claim 11, further comprising calculating the harmonic data by harmonic analysis of the ground fault current waveform data acquired from the simulated equipment. A current data acquisition unit that acquires ground fault current waveform data of a distribution line in which a ground fault has occurred,
A harmonic data calculation unit that calculates multivariate harmonic data indicating the intensity of harmonic components in the ground fault current waveform data by harmonic analysis of the ground fault current waveform data.
The harmonic data of the harmonic data is based on coordinate definition data calculated in advance by uncorrelated processing on a plurality of harmonic data so as to determine data coordinates for mapping the harmonic data of the ground fault current waveform data. A position calculation unit that calculates the position in the data coordinates,
Area data calculated in advance based on the correspondence between the plurality of harmonic data and the factor information of the plurality of ground faults and the harmonics so as to show the relationship between the position in the data coordinates and the factor of the ground fault. A ground relay that includes a factor estimation unit that estimates the cause of the ground fault based on the position of the wave data in the data coordinates.